Economic impacts of immigration detention reform - Cambridge … · 2019. 5. 8. · Economic impacts of immigration detention reform Cambridge Econometrics 3 Authorisation and version

Final Report March

2019

Cambridge Econometrics

Cambridge, UK

[email protected]

www.camecon.com

Liberty

Economic impacts of immigration detention reform

mailto:[email protected]

http://www.camecon.com/


2 Cambridge Econometrics

Cambridge Econometrics’ mission is to provide clear insights, based on rigorous and independent

economic analysis, to support policy-makers and strategic planners in government, civil society and

business in addressing the complex challenges facing society.

Cambridge Econometrics Limited is owned by a charitable body,

the Cambridge Trust for New Thinking in Economics.

www.neweconomicthinking.org

http://www.neweconomicthinking.org/





Authorisation and version history

Version Date Authorised for release by Description

1.1 08/03/19 Philip Summerton Client comments addressed.

1.0 27/02/19 Philip Summerton Draft report for comment.



Contents

Page

Executive Summary 5

1 Introduction 11

2 Approach 12

3 Results 13

4 Conclusion 32

5 References 34

Table 3.1: Days spent in detention by time band 14

Table 3.2: Detention days saved by time band 15

Table 3.3: Average daily cost of immigration detention by year 15

Table 3.4: Estimated cost savings under a 28-day time limit 22

Table 3.5: Detention days under a 28-day limit 23

Table 3.6: Estimated detention costs under a 28-day time limit 23

Table 3.7: Alternative estimated cost savings under a 28-day time limit 23

Table 3.8: Annual costs per participant of the Community Support Project 27

Table 3.9: Estimated cost of day replacement support beyond 28 days 28

Table 3.10: Estimated cost of year-long support beyond 28 days 29

Table 3.11: Estimated cost of year-long support to all people 29

Figure 3.1: Numbers of people released each year 18

Figure 3.2: Proportion of people held in detention by time band 18

Figure 3.3: Proportion of people released by outcome 19

Figure 3.4: Average daily cost of immigration detention by quarter 20

Figure 4.1: Savings and costs of UK immigration detention reform 32

Box 3.1: Data on average daily detention cost 16

Box 3.2: Defining Alternatives to Detention 26

Tables

Figures

Boxes



Executive Summary

This report presents the results of a study for Liberty on the potential costs

and savings of UK immigration detention reform. Specifically, the research

considers the economic impacts were the UK to introduce a 28-day time limit

on immigration detention. This contrasts with the current situation, under

which there is no time limit.

• As of March 2019, the UK is the only country in Europe without a statutory

time limit for holding someone in immigration detention. The UK also has

one of the largest detention estates in Europe.1 Pressure has been

mounting to introduce a statutory 28-day time limit, from a wide range of

organisations.

− In 2018, around 25,000 people entered detention and a similar number

left. Encouragingly, these figures are lower than the historical trend.

Over 2010-17, the number of people entering and leaving detention

ranged from 26,000 to 30,000 each year. The exception is 2015, when

the numbers climbed to 32,400 and 33,200, respectively.2,3

− On average, around 65% of people detained are released within 28

days. On the same basis, 95% are released within six months and 99%

within a year. Again, the 2018 figures show a possible departure from

trend, with a higher proportion of people released within 28 days

(69%).4

− However, Home Office statistics going back to 2010 also show

individual cases of people being held for much longer: over four years,

at times.5 Many people can also face multiple stints of detention and

re-detention, which is not reflected in Home Office statistics.

• The argument in favour of a time limit is fundamentally one of human rights

and there are recurring criticisms of the harm caused to people’s health

and wellbeing.6 These effects may persist beyond the period of detention,

especially in cases of prolonged detention.

In support of this, there may also be an economic case for a time limit, if

there are viable and cheaper alternatives to immigration detention. It is in

this context that Liberty commissioned Cambridge Econometrics (CE) to

examine the implications of a 28-day time limit.7

The UK is the

only country in

Europe without a

time limit on

immigration

detention

Immigration

detention is both

harmful and

costly: this

research focuses

on cost

Key findings

1 In 2017/18, Home Office expenditure on detention was £108m

2 Under a 28-day immigration detention time limit, there are potential

long-term cost savings of £55-65m each year

3 On the available evidence, plausible alternatives cost less than

detention, with a likely (albeit wide) cost range of £6-30m

A 28-day time limit supported by alternative provision could yield long-term

net savings of £25-35m or more each year. If sustained, a recent apparent

fall in the use of detention in 2018 would suggest cost savings closer to the

lower end, and to the clear benefit of those people who were either not

detained or released sooner.

https://www.libertyhumanrights.org.uk/

https://www.camecon.com/



Our analysis focuses on direct financial costs and benefits. It does not

consider, for example, the potential benefits of reduced physical or mental

harm. The analysis that follows only considers the long-term savings

following immigration detention reform. The available data preclude any

analysis of the potential transitional savings from any prospective reform.

• For an economic analysis of this kind, information on the existing UK

immigration detention system is quite sparse. Moreover, alternatives to

detention are still in the pilot phase, with limited evidence on how they

might operate at scale. Nevertheless, with the data that are available, it is

possible to estimate:

− Cost savings from detaining people for shorter lengths of time (implying

a smaller detention estate) and, consequently, avoiding compensation

payments for wrongful detention. This draws on data from the Home

Office.

− Additional costs of alternatives to detention (i.e. in the community) that

support people to resolve their cases. This uses data provided to us by

Detention Action on their Community Support Project.8

• Had there been a 28-day time limit in the past, our analysis suggests:

− Cost savings from shorter detention lengths and compensation

payments avoided of £55-65m, which can be interpreted broadly as an

annual saving.9 An alternative calculation method finds a similar range

of savings and sensitivity analysis supports a conclusion that potential

savings are in the tens of millions of pounds (a lower bound of around

£35m).

− There is significant uncertainty about the annual costs of alternative

provision at scale and our analysis suggests a wide range, from as little

as £6m to as much as £30m. That latter figure is, however, likely to be

near the upper limit of the plausible range.

A 28-day time

limit supported

by alternative

provision could

yield long-term

net savings of

£25-35m each

year

Executive Summary Table 1: The current UK immigration detention system

Value Period Comment

Home Office detention costs £108m 2017/18 Falling over time (£137m in 2014/15).

Excludes costs such as healthcare and

escorting people to/from detention

centres.

Home Office compensation

payments for wrongful detention

£3m 2017/18 Falling over time (£4m in 2014/15).

People entering detention 24,748 2018 Average 28,900 over 2010-17

(highs of 30,400-32,400 over 2013-15

and a previous low of 25,904 in 2010).

People leaving detention 25,487 2018 Average 28,900 over 2010-17

(highs of 30,000 in 2013 and 33,200 in

2015 and a previous low of 25,959 in

2010).

Average daily detention cost £87.71 2018Q4 Fell from £98.70 in 2014Q2 to £85.47 in

2017Q1. Rising from 2017Q1, until

2018Q4 when it fell slightly.

Source(s): Home Office (see Endnote 7).

https://detentionaction.org.uk/



Executive Summary Figure 1: Savings and costs of UK immigration detention reform

Source(s): CE calculations.

From our analysis of historical data on immigration detention and our

assessment of likely future policy developments, it is reasonable to think of

the implied net saving of at least £25-35m each year as indicative of the

future long-term impact. The net saving could only accrue gradually as

existing private contracts expire and alternative provision expands.

Without more detail on the breakdown of the Home Office’s expenditure, it

is difficult to make a like-for-like comparison but, for reference, Home

Office expenditure on detention was £108m in 2017/18.

It is difficult to conceive of any plausible situation in which the additional

costs of alternative provision could outweigh the estimated savings from a

28-day time limit.

• Limitations to our approach rest on data availability and quality:

− While there are regularly published statistics on some elements of the

system, including how long people are held for, information is limited

on what happens to those who are released into the community rather

than those who depart or are removed from the UK. This includes the

outcome of any claims for financial support that would ultimately be

paid by the UK government. It is not possible to construct firm

estimates of the costs of this support, owing to a lack of data about

how many people are or might be eligible, and whether they can

successfully claim that support. Low eligibility and low claims would

tend to lower the additional costs of support. However, any potential

financial saving here must also be understood as a decision about

what the state does or does not choose to provide to people so

affected. We assess the additional costs of financial support as

potentially low but with no clear way to establish an upper bound on

how large they could be. It was not possible to estimate the impacts on

public service provision.

− While Home Office annual reports and accounts do include detention

costs as a line item, this does not represent the entirety of the financial

costs associated with immigration detention.10 These figures do not

Limitations: data

availability and

quality



include, for example, the costs of healthcare provision (which are

borne separately, by NHS England), escorting services to and from

detention centres or the cost of legal cases settled out of court.11 In

some cases, the government withholds information for reasons of

commercial confidentiality.12 In this sense, there is scope for other

savings that are not captured in our estimates.

− Beyond the totals described above, and a Home Office-published

figure on average daily detention cost, there is little further detail on the

breakdown of operating costs.13 The average daily cost figures are

simply the operating costs of the detention centres divided by the

average number of bed spaces.14 Ideally, there would be more detail

available on, for example, fixed versus variable costs as well as a

breakdown of cost items. This is a critical uncertainty in our analysis

but, as a government-published estimate, it is the most credible figure

available on which to base our estimates.

− Alternative provision to reduce or avoid detention remains under-

developed in the UK.15 The Home Office has just started funding a

small pilot project to be delivered by the Action Foundation for women

who would otherwise be liable to detention at Yarl’s Wood. There is

also an independently funded Alternative to Detention, the Community

Support Project, run by Detention Action, which works with young men

with previous convictions who are liable to immigration detention.

These pilots are both small: their cost and effectiveness at scale is not

clear. In the case of the Community Support Project, on the one hand,

the data are for a programme of wider and longer-term support (around

one year per participant) than many might need. However, on the other

hand, the Community Support Project does not provide

accommodation, which may be a source of substantial cost in other

programmes. More research is needed to better understand any future

combination of prospective alternative provision.

While acknowledging the above as limitations, we see no reason to think

that the order of magnitude of our results is unreasonable.

• Uncertainties associated with our estimates concern the impacts of future

policies. In particular:

− The Immigration and Social Security Co-ordination (EU Withdrawal)

Bill, which proposes to change the definition of who is eligible to remain

in the UK. EU nationals make up an increasing proportion of those held

in immigration detention and the Bill could lead to more being

detained.16 Were this to happen, the potential savings from reduced

detention would likely increase.

− The implications of Brexit, which could change the number and

balance of EU and non-EU nationals coming to the UK, and about

which the cost impacts are unknown.

− The planned expansion of Heathrow Airport would require the

demolition of both the Harmondsworth and Colnbrook centres, with

new capacity expected to be built elsewhere.17 The implications of this

for future detention capacity and for potential future savings are

unclear.

Uncertainties:

future policy



Endnotes

1 Silverman, S.J., Griffiths, M. (2018) ‘Immigration Detention in the UK’, The Migration Observatory at the University of Oxford, Briefing, 03/07/2018 https://migrationobservatory.ox.ac.uk/resources/primers/immigration-detention-policy-challenges/

2 Home Office (2019) ‘Immigration statistics, year ending December 2018’, Detention: Table dt_01, ‘People entering detention by age, sex and place of initial detention‘ https://www.gov.uk/government/statistics/immigration-statistics-year-ending-december-2018

3 Home Office (2019) ‘Immigration statistics, year ending December 2018’, Detention: Table dt_06, ‘People leaving detention by reason, sex and length of detention’ https://www.gov.uk/government/statistics/immigration-statistics-year-ending-december-2018

4 Analysis of annual data over 2010-18 from: Home Office (2019) ‘Immigration statistics, year ending December 2018’, Detention: Table dt_06, ‘People leaving detention by reason, sex and length of detention’ https://www.gov.uk/government/statistics/immigration-statistics-year-ending-december-2018

5 Home Office (2019) ‘Immigration statistics, year ending December 2018’, Detention: Table dt_14_q ‘Top 20 longest lengths of detention of people in detention by sex’ https://www.gov.uk/government/publications/immigration-statistics-year-ending-september-2018/list-of-tables#detention

6 See, for example, the summary in: McGuinness, T., Gower, M. (2018) ‘Immigration detention in the UK: An overview’, House of Commons Library briefing paper, CBP-7294 https://researchbriefings.parliament.uk/ResearchBriefing/Summary/CBP-7294

7 Sources for Executive Summary Table 1:

a) Home Office detention costs: Home Office (no date) ‘Annual reports and accounts’ https://www.gov.uk/government/collections/ho-annual-reports-and-accounts

b) Home Office compensation payments for wrongful detention: Home Office (no date) ‘Annual reports and accounts’ https://www.gov.uk/government/collections/ho-annual-reports-and-accounts

c) People entering detention: Home Office (2019) ‘Immigration statistics, year ending December 2018’, Detention: Table dt_01 ‘People entering detention by age, sex and place of initial detention’ https://www.gov.uk/government/statistics/immigration-statistics-year-ending-december-2018

d) People leaving detention: Home Office (2019) ‘Immigration statistics, year ending December 2018’, Detention: Table dt_06 ‘People leaving detention by reason, sex and length of detention’ https://www.gov.uk/government/statistics/immigration-statistics-year-ending-december-2018

e) Average daily detention cost: Home Office (no date) ‘Migration transparency data’, Immigration enforcement: Table DT_02 ‘Average cost per day to hold an individual in immigration detention’ https://www.gov.uk/government/collections/migration-transparency-data#immigration-enforcement

8 Detention Action (no date) ‘Community Support Project’ https://detentionaction.org.uk/community-support-project/

https://migrationobservatory.ox.ac.uk/resources/primers/immigration-detention-policy-challenges/

https://migrationobservatory.ox.ac.uk/resources/primers/immigration-detention-policy-challenges/

https://www.gov.uk/government/statistics/immigration-statistics-year-ending-december-2018


https://www.gov.uk/government/publications/immigration-statistics-year-ending-september-2018/list-of-tables#detention


https://researchbriefings.parliament.uk/ResearchBriefing/Summary/CBP-7294

https://www.gov.uk/government/collections/ho-annual-reports-and-accounts






https://www.gov.uk/government/collections/migration-transparency-data#immigration-enforcement


https://detentionaction.org.uk/community-support-project/



9 Strictly, the analysis estimates the savings had everyone released in a calendar year (from the data) only been held for a maximum of 28 days. Some of those people would have consequently been released in earlier years with some savings falling in those earlier years. To the extent that the historical trend appears broadly stable, the saving can be thought of as approximately annual on the basis that, in a given year, while some savings would fall in earlier years, similar savings from later years would fall in the current year. This interpretation is clouded somewhat by the reduced use of detention in 2018 but the approximate order of magnitude in the results remains valid.

10 Home Office (no date) ‘Annual reports and accounts’ https://www.gov.uk/government/collections/ho-annual-reports-and-accounts

11 Hansard (2019) ‘Immigration: Detainees’, Response by Caroline Nokes to written question from Tulip Siddiq, 18/02/2019, 218833 https://www.parliament.uk/business/publications/written-questions-answers-statements/written-question/Commons/2019-02-08/218833

12 Hansard (2019) ‘Immigration: Detainees’, Response by Caroline Nokes to written question from Tulip Siddiq, 25/02/2019, 223185 https://www.parliament.uk/business/publications/written-questions-answers-statements/written-question/Commons/2019-02-19/223185/

13 Home Office (no date) ‘Migration transparency data’, Immigration enforcement: Table DT_02 ‘Average cost per day to hold an individual in immigration detention’ https://www.gov.uk/government/collections/migration-transparency-data#immigration-enforcement

14 Hansard (2019) ‘Immigration: Detainees’, Response by Caroline Nokes to written question from Tulip Siddiq, 01/02/2019, 212416 https://www.parliament.uk/business/publications/written-questions-answers-statements/written-question/Commons/2019-01-24/212416/

15 Detention Forum (2018) ‘Alternatives to detention: Frequently asked questions’ http://www.detentionforum.org.uk/wp-content/uploads/2018/08/AlternativestoDetention_A4_5.pdf

16 Analysis of annual data over 2010-17 from: Home Office (2018) ‘Immigration statistics, year ending September 2018’, Detention: Table dt_08, ‘People leaving detention by country of nationality, reason, sex and age’ https://www.gov.uk/government/publications/immigration-statistics-year-ending-september-2018/list-of-tables#detention

17 Shaw, S. (2018) ‘Assessment of government progress in implementing the report on the welfare in detention of vulnerable persons’ https://www.gov.uk/government/publications/welfare-in-detention-of-vulnerable-persons-review-progress-report


https://www.parliament.uk/business/publications/written-questions-answers-statements/written-question/Commons/2019-02-08/218833


https://www.parliament.uk/business/publications/written-questions-answers-statements/written-question/Commons/2019-02-19/223185/






http://www.detentionforum.org.uk/wp-content/uploads/2018/08/AlternativestoDetention_A4_5.pdf




https://www.gov.uk/government/publications/welfare-in-detention-of-vulnerable-persons-review-progress-report




1 Introduction

This report presents the results of a study for Liberty on the potential costs

and savings of UK immigration detention reform. Specifically, the research

considers the economic impacts were the UK to introduce a 28-day time limit.

This contrasts with the current situation, under which there is no time limit.

As of March 2019, the UK is the only country in Europe without a statutory

time limit for holding someone in immigration detention. Moreover, the UK has

one of the largest detention estates in Europe (Silverman and Griffiths, 2018).

Pressure has been mounting to introduce a statutory 28-day time limit, from a

wide range of organisations.

The argument in favour of a time limit is fundamentally one of human rights

and there are recurring criticisms of the harm caused to people’s health and

wellbeing (see, for example, McGuinness and Gower, 2018). These effects

may persist beyond the period of detention, especially in cases of prolonged

detention.

In support of this, there may also be an economic case for a time limit, if there

are viable and cheaper alternatives to immigration detention. It is in this

context that Liberty commissioned Cambridge Econometrics to examine the

implications of a 28-day time limit.

The rest of this report is structured as follows:

• Chapter 2 sets out our approach

• Chapter 3 presents the analysis itself, with results divided into sections on

cost savings and additional costs incurred (from the provision of

alternatives to detention)

• the report concludes in Chapter 4

We provide references in Chapter 5.



2 Approach

The aim of the research is to analyse how the balance of costs and benefits

might differ under a 28-day time limit by considering:

• Potential cost savings from holding people for shorter lengths of time and,

by implication, running a proportionally smaller detention estate.

− Alongside that, we account for further savings to the Home Office under the assumption that a 28-day time limit would avoid all compensation payments for wrongful detention. As we explain in the next chapter, our overall conclusions are not sensitive to this assumption.

− As we discuss later, the cost savings would ideally also include savings from reduced costs elsewhere as a result of reduced detention e.g. healthcare and escorting to/from detention centres. However, there were not enough data to inform such an analysis.

• Likely additional costs incurred from the need to provide alternative

provision in place of detention.

− For this, we consider the costs of providing casework support to people in the community, as an alternative to detention. The figures we use here are based on information provided by Detention Action about their Community Support Project.1

− In principle, this part of the analysis should also include the costs of support that people might be eligible for outside of detention. While there is information on the level of support that people can claim, there is little evidence on how many people successfully claim that support and for how long. We discuss this further in the next chapter.

The difference between the first (savings) and the second (new costs) gives

the net saving or cost of a possible immigration detention reform. Our

approach is to carry out this analysis using past data and to then examine the

assumptions under which these results can be interpreted as appropriately

representative of potential future impacts i.e. that they can be extrapolated

reasonably.

Broadly, our approach mirrors that of an earlier study by Matrix Evidence

(2012) for Detention Action. This current research differs by:

• considering a firm 28-day time limit, in contrast to the earlier report’s

assessment of timely release after three months, in cases in which people

would eventually be released back into the community anyway

• making use of more recent data on the structure and operation of the UK

immigration detention estate, up to 2018, whereas the previous study used

data up to 2010

In the next chapter, we set out our approach in more detail, alongside the

results of our analysis. This includes an assessment of the applicability of

these estimates as potential future cost savings.

1 https://detentionaction.org.uk/community-support-project/




3 Results

This chapter presents, in separate sections, the results of our analyses of cost

savings and additional costs incurred. In each case, we set out our approach

and underlying assumptions before presenting the results and discussing any

limitations.

3.1 Cost savings

This section presents our estimates of the potential cost savings were there to

be a 28-day time limit on immigration detention. From our analysis of Home

Office data and UK government information, we conclude that there are

potential savings of £55-65m each year under a 28-day time limit. Our results

and assessment of recent trends in the operation of the detention estate

suggest potential cost savings towards the lower end of that range if the 2018

figures are taken as indicative of the long-term trend. However, if this is the

case, it is most likely because the Home Office is, effectively, already

beginning to realise some cost savings of using detention less. More

discussion of this follows. We also present an alternative calculation that

yields similar results to our principal estimates.

The main uncertainties associated with the above figure concern:

• the quality of the UK immigration detention data, including on costs which

could not be included in our analysis

• the extent to which the current condition of the detention estate is a

reliable indicator of its future operation, should indefinite detention

continue

• the possible impacts of legislative and policy developments that are still to

be resolved at the time of writing, including the Immigration and Social

Security Co-ordination (EU Withdrawal) Bill 2017-19 and Brexit

In the following sections we present our approach and its underlying

assumptions, the results of the analysis and a discussion of the reliability of

our results.

Approach

Using data for 2014-18, our approach is to estimate the cost savings from

implementing a 28-day time limit by multiplying:

1 The total number of detention days saved were there to be a time limit –

the difference between the total number of days that detainees are held

under the current system and the total number of days they might be held

under a 28-day system.

2 The average cost per day to hold someone in detention.

To this we add the potential savings from compensation payments avoided

for wrongful detention. We do this for the years 2014-17. We explain each of

these components below.

Note that, owing to a lack of data, these estimates exclude other costs

associated with immigration detention such as the cost of escorting people to

and from detention, and healthcare provision (Hansard, 2019b). In these

We focus on the impacts of

reduced detention on

detention-centre and legal costs



respects, what can be captured in the analysis is somewhat narrow in scope.

Conversely, and other things being equal, these costs can be thought of as

being towards the lower end of the cost savings that might be achieved under

a 28-day time limit.

The Home Office publishes data on the number of people released from

detention, broken down by the length of time before release, in days.2 These

data are reported in bands e.g. ‘3 days or less’, ‘4 to 7 days’ etc. We do not

know precisely how long people are held in detention before release. The size

of the bands, and in turn the accompanying uncertainty, increases with longer

detention lengths.

Because the statistics provide detention lengths as ranges, we do not know if

someone who appears in the category ‘18 months to less than 24 months’

spent closer to 548 or 729 days in detention. This is a material difference in

both human and financial terms. Given this uncertainty, in our analysis we

consider three alternative assumptions about the typical detention length in

each category: lower, central and upper estimates. The lower estimate is the

minimum that someone could spend in detention in each category, while the

upper estimate is the maximum. The central estimate is the midpoint between

the two (see Table 3.1).

Table 3.1: Days spent in detention by time band

Detention length Lower Central Upper

3 days or less 0.0 1.5 3.0

4 to 7 days 4.0 5.5 7.0

8 to 14 days 8.0 11.0 14.0

15 to 28 days 15.0 21.5 28.0

29 days to less than 2 months 29.0 44.5 60.0

2 months to less than 3 months 61.0 75.5 90.0








48 months or more 1460.0 1460.0 1460.0


Note that in Table 3.1 we assume that someone in the longest detention

length category (’48 months or more’) is held for no more than 48 months (four

years). Rather than assume central and upper limits on detention lengths, we

have been conservative by applying a common lower limit.3 In any case, the

2 Home Office ‘Immigration statistics quarterly release’, Detention data tables, Table DT_06

https://www.gov.uk/government/collections/immigration-statistics-quarterly-release

3 Longer assumed detention lengths under current conditions would create greater opportunities for savings.

By assuming the shortest possible detention length for this category, our final cost estimates are at the

lower end.

Total number of detention days

saved

https://www.gov.uk/government/collections/immigration-statistics-quarterly-release



number of releases in this category accounts for less than 0.1% of those

released from detention and thus has a negligible impact on the results

derived from these data.

From the assumptions in Table 3.1, we can then estimate the number of

detention days saved (reduced) per person under a 28-day limit. This is

calculated by subtracting 28 days from the numbers in Table 3.1. For the

bands that have an upper limit of 28 days or less, no detention days are

saved. We assume that the implementation of a 28-day limit would not have

an effect in terms of cost savings (see Table 3.2).

Table 3.2: Detention days saved by time band

Detention length Lower Central Upper

3 days or less 0.0 0.0 0.0

4 to 7 days 0.0 0.0 0.0

8 to 14 days 0.0 0.0 0.0

15 to 28 days 0.0 0.0 0.0

29 days to less than 2 months 1.0 16.5 32.0









48 months or more 1432.0 1432.0 1432.0


We then multiply the numbers of detention days saved per person per band by

the total number of people in each band to derive the total number of detention

days saved under a 28-day detention limit. This is the first component of the

cost savings calculation.

To get a total cost saving estimate, we multiply the total number of detention

days saved by an average cost per day to hold someone in detention. This

average cost figure is provided quarterly by the Home Office as part of the

‘Immigration enforcement transparency dataset’. We calculate the annual

averages as the means of these quarterly data (see Table 3.3). This gives the

second component of the cost savings calculation.

Table 3.3: Average daily cost of immigration detention by year

Year Average cost per day per person (£)

2014 97.32

2015 91.32

2016 88.03

2017 85.72

2018 87.61

Source(s): Home Office (2019b); CE calculations.

Cost per day



Total cost savings then follow by multiplying the number of detention days

saved by the average daily cost. This estimate gives an answer to the

following question:

Of the people who were released from immigration detention

in a particular year, what are the total costs that might have

been saved had there instead been a 28-day time limit for

these people?

Strictly, this is not the same as an annual cost savings figure because not

everyone released in a specific year was held solely in that year. For example,

someone released in 2017 after spending between 18 and 24 months in

detention would also have spent some of 2016 in detention, and possibly

2015. The same applies, albeit to a somewhat lesser extent, to shorter

detention lengths. Someone released after 2-3 months could have spent some

of 2016 in detention, had they been released in January or February 2017.

Nevertheless, as we explain later, under certain conditions, the results can be

interpreted broadly as an annual figure.

Box 3.1: Data on average daily detention cost

Each quarter, the Home Office publishes, as part of its Immigration

Enforcement data release, a figure on the ‘average cost per day to hold an

individual in immigration detention’ (in Table DT_02).

The Home Office derives this figure by ‘dividing the Total Resource Costs

of running the Detention Centres (Contracts, Staff, Rent, Rates, Utilities

and Depreciation) by the average number of bed spaces’ (Hansard,

2019a). Consequently, and as we show in this section, the average daily

detention cost figure varies from quarter to quarter.

The figure gives an indication of the average cost to the taxpayer, that

quarter, of a detention day given running costs and the capacity of the

detention estate. It may vary for reasons other than changes in the actual

operating cost of detaining someone for a single day and/or according to

the capacity of the detention estate at that time.

While not the same as the underlying cost of holding someone for an

additional day, the figure serves our purpose by representing the cost to

the government (and by extension, the taxpayer) of a detention day that

quarter. It is also the only readily available figure on per-day detention

costs. To the extent that it is the government’s own figure, it is the best

available estimate for our analysis.

Moreover, the average cost figure does not include associated costs of

detention such as the process to determine someone’s right to stay in the

UK (Legal Aid and Home Office Administration), the cost of escorting them

or any health care (Hansard, 2019b). These other costs could be affected

by the implementation of a 28-day time limit but are not included in our

analysis owing to a lack of detailed data. Some of these data have not

been disclosed for reasons of commercial confidentiality (see, for example,

Hansard, 2019c). We discuss this limitation later in this section.



To the above cost savings, we add in further savings from compensation

payments avoided for wrongful detention. In 2014/15, the Home Office paid

out £4m in compensation, falling to £3m in 2017/18. As we discuss later, we

assume that these costs could be entirely avoided under a 28-day time limit.

The next section discusses recent trends in the detention estate and their

implications for our modelling assumptions and the validity of the cost

analysis.

Assumptions

The purpose of the analysis is to identify the new balance of costs and savings

under a 28-day time limit for immigration detention. This exercise must

extrapolate results into the future to give an indication of that balance of costs

and savings were a time limit to be introduced. With only historical data to go

on, the validity of the results from that forward-looking exercise depends on

what we assume about the detention estate under a ‘business as usual’ case

(indefinite detention) and how that might change under a 28-day time limit.

In this section we examine recent historical trends in detention (over 2014-18)

and consider what this might mean for the future and potential long-term cost

impacts. Specifically, we ask whether:

1 The detention estate holds a similar number of people each year; it is

neither growing nor contracting markedly.

2 The detention estate releases the same proportion of people each year, by

length of detention.

3 The detention estate releases the same proportion of people each year, by

reason for release.

4 The average cost per day of holding someone in detention is constant over

time.

If the total number of people released from detention each year were to

change significantly in the future, then the potential cost savings would also

change. For example, if the future trend were for an increase in the number of

people held, there would be greater scope for cost savings under a 28-day

time limit.

As Figure 3.1 shows, in most years, between 26,000 and 30,000 people were

released from immigration detention. The mean over 2010-17 is around

29,000 releases each year. There is a clear outlier in 2015, when the

‘Detained Fast Track’ system ended after being deemed unlawful by the High

Court (McGuinness and Gower, 2018). In that year, some 33,200 people were

released.

The number of releases in the last year of data, 2018, appears unusually low

compared to the previous six years, with around 25,500 people released. The

numbers of people entering detention (not shown) is similarly low compared to

previous years. It is not yet clear whether this fall is temporary or not. If it

reflects a sustained fall, the scope for future cost savings will be reduced but,

crucially, because the use of detention will have fallen. This is beneficial in

human terms and can also be thought of as the Home Office already realising

some savings from reduced use of detention.

A development not captured in the data available during this study is the

closure of Campsfield House. Campsfield House stopped holding people in

Compensation payments

avoided for wrongful detention

The number of people released

from detention each year



December 2018 and, prior to closure, represented around 9% of the detention

estate’s capacity. What this means for future trends is not yet clear. On the

one hand, a smaller detention estate may lead to fewer people being detained

than previously. On the other hand, capacity utilisation may increase such that

the same number of people are detained each year. In our analysis, we adopt

the latter assumption. We discuss the implications of this later in this section.

Figure 3.1: Numbers of people released each year

Source(s): Home Office (2019a).

To calculate the number of detention days saved, we need to know both the

number of people released (as above) and how long they spent in detention. If

a trend was observed in the data, such as a rising proportion of people kept

for more than six months, this would have implications for the number of days

saved under a 28-day time limit. In this case, the total number of days saved

in the future would increase, leading to greater scope for savings.

Figure 3.2: Proportion of people held in detention by time band


0

5,000

10,000

15,000

20,000

25,000

30,000

35,000

2010 2011 2012 2013 2014 2015 2016 2017 2018

Up to and including 28

days

29 days - 2 months

2 - 3 months

3 - 4 months

4 - 6 months

6 - 12 months

12+ months

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2010 2011 2012 2013 2014 2015 2016 2017 2018

The proportion of people

released from detention each year, by length

of detention



Home Office data indicate that the proportions of people held for different

lengths of time have been approximately constant for some time (see Figure

3.2). Since 2010, around 65% of people have been held for less than 28 days.

The proportions of people held between 28 days and 2 months, and 2 and 3

months, are similarly constant, at around 15-20% and 5-10% respectively.

However, in 2018, there was a slight increase in the proportion of people

released within 28 days. As with the numbers of people released (above), this

points to some reduction in the use of detention in 2018. We do not currently

know if this is a sustained change or not. Longer detention-length bands also

appear relatively stable over time though there is more scope for variation

because the numbers of people are much smaller in these categories.

Whilst data on the costs of detention broken down by the reason for release

are not available, trends in detention outcomes are relevant because those

costs may differ between, for example, someone released on bail and

someone removed from the UK. While not captured in the average cost per

day figure to hold someone in detention (see Box 3.1), this shifting

composition may signal underlying developments that affect the average cost

figure.

Unlike the previous graphs, there is a trend in these figures, with no clear sign

as to whether it will continue or level off (see Figure 3.3). The proportion of

people released on bail by the Secretary of State and by Immigration Judges

has increased since 2010. The proportion of people ‘returned from the UK’ has

correspondingly decreased. Without more information, it is not possible to

know what impact this might have on the average cost in the future.

Figure 3.3: Proportion of people released by outcome


Each quarter, Home Office immigration enforcement transparency data

provide a daily average cost of holding someone in immigration detention,

from 2014Q2 onwards. As Figure 3.4 shows, the cost fell significantly between

2014Q2 and 2017Q1, reaching a low of £85.47 per detention day.4 Since then

the cost has been steadily rising, to £88.29 in 2018Q3, before falling slightly,

4 This compares to an even higher cost of £110 per day as used in the earlier Matrix Evidence (2012) study.

Bailed -Immigration

Judge

Bailed -Secretary of

State

Granted Leave to Remain /

Enter

Returned from the UK

Other

0%

20%

40%

60%

80%

100%

2010 2011 2012 2013 2014 2015 2016 2017 2018

The proportion of people

released from detention each year, by reason

for release

The average cost per day of

holding someone in

detention



to £87.71 in 2018Q4. Whilst the average cost of detention per detainee day

has not been stable over the period for which there are data, the recent

upward trend has been mild.

Until 2018Q4 the average cost had not fallen since 2017. It is not clear what

the future trend might be, therefore we make no further assumption about how

costs may change over time, either up or down.

As above, the closure of Campsfield House (Home Office, 2018b) may also

have implications for the future trend. The closure simultaneously reduces the

capacity of the detention estate and its total cost. How this might affect the

average daily cost depends on how Campsfield House’s average cost

compared to other detention centres. Again, in the absence of more detailed

information, we have no strong basis on which to form an alternative

assumption.

Figure 3.4: Average daily cost of immigration detention by quarter

Source(s): Home Office (2019b).

The evidence presented in this section suggests that:

1 Barring a spike in 2015, the number of people detained over the course of

a year had been relatively stable over 2010-17, at around 29,000 people.

However, the number fell substantially in 2018 and it is not clear what this

might mean for the future trend.

− There is also some uncertainty as to how the closure of Campsfield House might affect these figures but here we assume that the rest of the estate will accommodate the additional people. Though we cannot be sure, it is possible that this overstates the number of detention days saved in our calculation.

2 The proportion of people held by different detention lengths has been

broadly stable since 2010, with around 35% of those released having been

held for longer than 28 days. In 2018, this proportion fell to 31%.

3 The pattern of detention outcomes has not been stable over time, with

people increasingly released into the community rather than returned from

the UK.

− With the data available to us, this has no bearing on our analysis. We are left to assume that the costs (and any associated costs) are similar.

75

80

85

90

95

100

2014 2015 2016 2017 2018

£

Past trends have proven relatively

stable, but it is not yet clear how

the 2018 figures will affect the

future trend



4 The average daily cost of detention has not been stable, first falling and

then rising from 2017 onwards. This remained the case until 2018Q4,

when the cost fell slightly.

− A gradual upward trend in costs is evident in the data; nevertheless, we opt to use the historical figures, in lieu of attempting to predict the trend. If, after the fall in 2018Q4, costs were to continue to increase in the future, our calculation will underestimate the cost savings.

Results

Using the available data over 2014-18, our historical analysis suggests that a

28-day time limit on immigration detention could have saved £55-65m in

each of those years. Given the trends (or otherwise) examined in the

previous section, we think that it is reasonable to think that similar long-

term savings could be achieved in the future. The upper and lower bounds

indicate that this estimate lies in the range £35-90m (see Table 3.4). Even at

the lower bound, the scope for savings is in the tens of millions of

pounds. Given the uncertainties in the source data, we recommend that these

figures be presented and discussed in terms of the nearest £5m or £10m.

Strictly, these savings cannot be directly interpreted as an annual cost saving.

Rather, they represent the savings that could have been made had the people

released each year been held for no more than 28 days. Those savings could

have been realised in earlier years. In that respect, they represent a ‘lifetime’

saving associated with those who left detention in a particular year.

Nevertheless, to the extent that the historical trend appears broadly stable

(from the previous section), the saving can be thought of as approximately

annual on the basis that, in a given year, while some savings would fall in

earlier years, similar savings from later years would fall in the current year. If

the 2018 fall persists the potential cost savings would be at the lower end.

To give some sense of scale, the annual cost to the Home Office of detention,

as identified in its annual accounts, was just over £108m in 2017/18 (Home

Office, 2018c). Without knowing more about the underlying components of the

Home Office’s detention expenditure, we cannot be sure that the comparison

is like-for-like. Regardless, the estimated cost saving is certainly material

when set against the Home Office figure. If associated costs (e.g. healthcare

and escorting) were also to fall, then further savings could be made

elsewhere.

As well as the savings from holding people for shorter lengths of time, we also

include the savings from avoiding compensation payments for wrongful

detention. In Table 3.4 we assume that these compensation costs can be

completely avoided. We assume this on the basis that people held for no more

than 28 days are less likely to have been detained in a way that would breach

the Hardial Singh principles for detention, as a claim would have to satisfy that

they had been detained for an unreasonable period given all the

circumstances. Without satisfying all the legal principles, they would thus be

unable to claim compensation (ILPA, 2011). This has been included in the

second-to-last column of Table 3.4.

Compensation payments have fallen over time and currently represent an

extra 5% in savings over the core cost saving figure. Whether these costs

would genuinely be avoided has no bearing on the overarching message from

A 28-day time limit could

generate £55-65m in cost

savings

We include

savings from

compensation

payments, but

these are small



these results. Were these costs still incurred, our cost savings estimate would

be around £3m less against an overall saving of £55m or more. Note that

Table 3.4 excludes compensation payments for 2018 as the corresponding

Home Office accounts have not yet been published. Consequently, the saving

for this year is likely to be higher than the figure of £54.4m in the table.

These figures exclude the cost of legal cases settled out of court, which could

be a source of further potential savings under a 28-day time limit.

The total savings figure for 2015 stands out due to its particularly high central

estimate of cost savings: over £80m compared to the other figures, which lie in

the £55-65m range. This is due to the particularly high number of detainees

released in that year and possibly related to the end of the Detained Fast-

Track system.

Table 3.4: Estimated cost savings under a 28-day time limit

Detention

days saved

Average daily

cost (£)

Cost saving (£m) Compensation for wrongful

detention (£m)

Estimated saving (£m) Lower Central Upper

2014 700,866 97.32 44.1 68.2 92.3 4.0 68.2 [44.1, 92.3]

2015 885,894 91.32 52.9 80.9 108.9 4.1 80.9 [52.9, 108.9]

2016 672,241 88.03 38.0 59.2 80.4 3.3 59.2 [38.0, 80.4]

2017 767,215 85.72 43.1 65.8 88.4 3.0 65.8 [43.1, 88.4]

2018 620,807 87.61 35.8 54.4 73.0 No data 54.4 [35.8,73.0]

Note(s): Cost savings figures are by calendar year while compensation figures are by financial year. For simplicity, we have taken the figures directly and not attempted to adjust for different year definitions. Lower and upper estimates of the total ‘Estimated saving (£m)’ are given in square brackets.

Source(s): Home Office (2019a, 2019b, 2018c); CE calculations.

In the analysis above, we calculated the costs saved had there been a 28-day

time limit, but it is also possible to estimate the costs of running the detention

estate under a 28-day time limit i.e. the number of detention days required

under a 28-day time limit, rather than the number of detention days saved. We

do this to test the robustness of our main results.

Here, we calculate the total number of days that would have been spent in

detention under a 28-day time limit and multiply the figure (as before) by the

average daily cost. The total number of days is calculated in a similar way to

before, using the banded data.

First, the total number of people entering the detention system each year is

split by the proportions leaving immigration detention that same year.5 We

then multiply these figures by the number of days spent in detention, with

lower, central and upper bounds as before. For those that were detained for

fewer than 28 days, we again assume that the implementation of the 28-day

limit would have had no financial effect. For all those previously held for more

5 The correspondence here is only approximate because not everyone released from immigration detention

was originally detained in the same year. Nevertheless, as the earlier section showed, the proportions are

quite stable over time.

We carry out an alternative

calculation as a robustness

check



than 28 days, the maximum time they would now be held falls to 28 days (see

Table 3.5).

Table 3.5: Detention days under a 28-day limit

Detention Length Lower Central Upper

3 days or less 0.0 1.5 3.0

4 to 7 days 4.0 5.5 7.0

8 to 14 days 8.0 11.0 14.0

15 to 28 days 15.0 21.5 28.0

29 days or more 28.0 28.0 28.0


The total number of days spent in detention under the 28-day limit, in a given

year, is then multiplied by the average daily cost. The resulting central cost

figure for the detention estate under the 28-day limit is in the range £33-48m

(see Table 3.6).

Table 3.6: Estimated detention costs under a 28-day time limit

Cost under 28-day limit (£m) Compensation for wrongful detention

(£m)

Average daily cost of

detention (£) Lower Central Upper

2014 40.4 45.9 51.3 4.0 97.32

2015 42.2 47.8 53.4 4.1 91.32

2016 35.3 40.2 45.2 3.3 88.03

2017 33.2 37.9 42.6 3.0 85.72

2018 27.4 31.9 36.5 No data 87.61


We then calculate an implied cost savings figure by subtracting the costs in

Table 3.7 from the Home Office expenditure figures on detention cost (see

Table 3.7). We do not report estimates for 2018 because the corresponding

Home Office accounts have not been published.

Table 3.7: Alternative estimated cost savings under a 28-day time limit

Cost saving (£m) Total detention cost (£m)

[Home Office accounts]

Upper Central Lower

2014 85.6 91.0 96.5 136.9

2015 71.6 77.2 82.8 125.0

2016 72.9 77.9 82.8 118.1

2017 65.4 70.1 74.8 108.0

Note(s): No figures for 2018 available. The corresponding Home Office accounts have not yet been published.


From 2015 onwards, the central figures for the cost savings are in the region

of £70-80m. This is close to (slightly higher than) the main cost saving

estimate for these years. The implied cost saving in 2014 is markedly higher,

at over £90m. Insofar as 2014 is less recent and the total costs of detention

were higher back then, we consider the later figures to be more representative

of the potential cost savings.



Limitations

The availability and quality of data on immigration detention is quite limited,

making it a challenge to assess a possible 28-day time limit at a high level of

detail. Nevertheless, our approach produces a cost saving figure that we find

plausible.

By excluding associated costs of detention (e.g. healthcare and escorting; see

Hansard, 2019b), our estimate is perhaps narrow in focus because of limited

data on these other aspects. To the extent that our cost savings exclude

elements that ought to be reduced by shorter detention lengths, there is likely

to be somewhat greater scope for savings than our figures first imply. It is also

possible that these changes might alter the need for other services like legal

aid. In lieu of detailed data to inform an analysis, we do not include effects like

these but note that some support, like legal aid, has been heavily cut over

time. This limits the scope for costs or savings either way.

An important factor to consider when analysing the possible cost savings from

a 28-day limit is the current structure of the detention estate. Currently, the

Home Office contracts out the operation of most detention centres to firms

such as Serco and G4S. These contracts have expiry dates moving into the

medium-term. As the government is legally obliged to fulfil these contracts it

will not be possible for the government to immediately achieve the cost

savings detailed above. There will be transitional costs of reducing the size of

the detention estate, but these are not captured in the analysis above.6

Instead, these cost savings should be considered as the long-term savings

that might be achieved from a proportionally smaller detention estate should

there be a 28-day time limit on immigration detention.

As mentioned above, another caveat to the analysis is that there is an implicit

assumption that the cost of detention per day is the same for all people by

detention length, eventual outcome and detention centre. Without further

information, it is difficult to know what impact this may have on the final cost

figures. Similarly, it is not currently possible to separate detention costs into

their fixed (e.g. rent and rates) and variable (per person) elements.

Consequently, and in line with the government data on average daily costs,

we treat the cost as a fully variable figure. The underlying assumption here is

that the cost of the detention estate under a 28-day limit is proportional to the

change in the number of people detained / detention days. While not

unreasonable to give a sense of the scale of the savings, without an itemised

breakdown of the spending required to run the detention estate (from the

Home Office), a more detailed model of a future alternative detention estate

cannot easily be constructed.

There is also some uncertainty as to the effects that Brexit might have on the

detention system. The Immigration and Social Security Co-ordination (EU

Withdrawal) Bill currently progressing through Parliament, has the potential to

affect the number of people who are detained because it increases the

number of people liable to the automatic deportations section of the UK

Borders Act 2007 (HM Government, 2007 and 2019). An increasing number of

EU nationals were already being held in the detention estate: if this trend were

6 To do this we would need information on the nature of the contracts, the Home Office does not disclose

this for reasons of commercial confidentiality.



to continue then the scope for cost savings would likely increase because

more people could be held for longer. This does, however, need to be

weighed up against the uncertainty of migrant numbers coming to the UK

post-Brexit.

More broadly, the validity of our estimates as an indication of future cost

savings depends crucially on future trends in the detention estate (as

examined earlier in this section). As we highlighted, the historical data mostly

suggest stable levels of detention until 2018 while the recent closure of

Campsfield House does not yet show up fully in the figures. If the use of

detention were to fall in line with this lower capacity (rather than increasing

capacity utilisation elsewhere), the estimated cost saving would be lower. A

proportionate fall in the use of detention would suggest a reduction in the

estimated cost saving of around 9%, other things being equal.

Furthermore, if the dip in 2018 in the number of people leaving detention

combined with an increasing proportion being held for 28 days or less

continues, the potential cost savings will be lower. We have no information

that indicates whether this change will continue but, if it does, the implication is

that reductions in the use of detention are already underway. A reduction in

the scope for cost savings of this kind suggests that cost savings are already

being realised.

However, we also note that the average daily cost of detention has generally

been rising. While (from Box 3.1) this cannot be directly interpreted as an

increase in the cost of detention (it is, in part, a reflection of how operating

costs are spread across detained people, on average), if this does represent

at least some trend increase in the underlying costs over time, then this

increases the scope for cost savings.

3.2 Additional costs incurred

The previous section set out our estimates of the potential annual savings of a

28-day time limit, entailing lower detention costs and a smaller detention

estate. However, viable immigration detention reform must also consider what

happens to people who would otherwise have been held in immigration

detention for longer than 28 days. There must be some consideration of the

accompanying costs of providing for people outside of detention i.e. in the

community. In this section we present our estimates of additional costs that

might be incurred as part of wholesale immigration detention reform.

Alternative provision to reduce or avoid detention remains under-developed in

the UK with some attempts to run pilot programmes in the 2000s (Detention

Forum, 2018). Detention Action has run a pilot Alternative to Detention, the

Community Support Project (CSP), since 2014.7 The Home Office has just

started funding a pilot project for women otherwise liable to detention in Yarl’s

Wood (Nokes, 2018).

7 https://detentionaction.org.uk/community-support-project/

Additional costs of a 28-day time limit arise from

the need for alternative provision




Detention Action provided us with information on the CSP to inform an

analysis of how much one form of provision might cost if provided as an

alternative to long-term detention.

Approach

The CSP is a scheme for young men (aged 18-30) who have completed

prison sentences and have either experienced or are at risk of long-term

immigration detention. The scheme provides support to participants to help

them comply with the terms of their release and avoid re-offending.

Using data provided to us by Detention Action, we calculate the average cost

per participant of running the CSP. We then calculate the total cost incurred of

scaling up the programme by multiplying the average cost so calculated by the

number of people and/or days. The CSP will support the typical participant for

around a year but many people in immigration detention spend much less than

a year in detention. We examine alternative support times to gauge the

approximate scale of cost that might be incurred.

Detention Action is currently hoping to expand its programme by adding two

additional caseworkers (compared to just one now) situated around the

country. This will allow the CSP to accept a larger number of participants and

for caseworkers to be based closer to their clients. In the analysis below, we

include an assessment that makes use of these additional figures.

Assumptions

Pilot programmes to provide alternatives to detention are both:

• small in scale, by virtue of being pilots

• typically designed to meet the needs of a specific group of people e.g. in

the case of the CSP, young men with previous convictions who need long-

term support

This introduces much uncertainty as to whether and how these pilots might

scale and contribute to a more comprehensive system of alternative provision

that caters for all people at risk of detention. While it is an unknown, we do

note that the CSP has shown enough promise for Shaw (2018) to include its

expansion as one of his recommendations to government.

Under the CSP, a project coordinator develops a transition plan with the

participant, meeting regularly to provide support. The CSP may also pay for a

participant’s gym membership. Typically, this support runs for a year, but this

may vary. Currently the CSP employs one caseworker with core costs

We use data on a live pilot

programme, run by Detention

Action

There are few live pilot

programmes and they are small in

scale

Box 3.2: Defining Alternatives to Detention

As the Detention Forum (2018) notes, there is no ‘universally established

legal definition’ (Page 2) of Alternatives to Detention. Moreover, there is not

necessarily agreement as to what constitutes an acceptable alternative.

In the context of this research, we have considered an alternative that is

compatible with principles that Liberty considers to be appropriate.

Specifically, a programme that is community-based and non-coercive in

nature, drawing on social work principles; in this case, Detention Action’s

Community Support Project.



comprising salary, rent, management and overheads. Currently, this amounts

to £45,000 each year with around 25 active participants at any one time.

Per participant, the principal costs are the caseworker’s travel and subsistence

costs. Meetings may cover the participant’s travel and refreshments. An

annual gym membership may also be paid for.

Detention Action provided cost data for a sample of five participants

representing a range of levels of cost/support (see Table 3.8). This gives an

average annual cost per participant of £963.41 or £24,085 for 25 participants

for a year. Combined with the fixed cost of £45,000, the implied total annual

cost of the programme is £69,085 or £2,763 per participant per year.

Table 3.8: Annual costs per participant of the Community Support Project

CLIENT

MEAN 1 2 3 4 5

Travel:

Caseworker (£)

100.00 3.00 5.80 50.00 90.00 49.76

Travel: Client (£) 4.00 4.00

Refreshments (£) 5.00 5.00 5.00 5.00

Lunch (£) 5.00 5.00 4.50 4.83

Total per visit (£) 110.00 8.00 10.80 59.00 94.50 56.46

Occurrences each year

12 26 26 26 9 19.8

Annual cost (£) 1320.00 208.00 280.80 1534.00 850.50 838.66

Annual gym membership

203.88 419.88 311.88

TOTAL 1320.00 208.00 484.68 1953.88 850.50 963.41

Source(s): Detention Action; CE calculations.

The implied average daily cost of the CSP is £7.57 with a range of

£5.50-£10.28. This compares to an average daily detention cost from the

previous section of almost £90. On a per-day basis, this form of alternative

provision is around one-tenth the cost of immigration detention.

In the event of an expansion, Detention Action also provided figures on the

additional (fixed) cost of two extra case workers, of £102,440, to cater for an

additional 40 people. Assuming the same average annual cost per participant,

the average additional cost of catering for 40 more participants is higher than

the existing cost: £3,524 compared to the current cost of £2,763. Detention

Action expects this increase in average cost in the early scaling phase but for

economies of scale to take effect were provision to be scaled further.

Moreover, the additional case workers would be located elsewhere in the

country, to more easily travel to certain participants. In that respect, the travel

costs (which can be a substantial proportion of total costs, as in Table 3.8)

may also fall. In the analysis that follows we present both the ‘current’

(original) and ‘additional’ (planned) cost estimates but consider the ‘current’

figures to be more likely to represent the total costs at scale. Given the point

about travel costs, it is possible that the average cost would be lower still, but

we do not consider it further in the analysis that follows.

A further argument is that, under a 28-day time limit, there is less potential for

harm to physical or mental health and that there is a greater likelihood of

On a per-day

basis, the

existing pilot

programme is

substantially

cheaper than

immigration

detention



someone coming from or returning to a stable environment in the community.

In time, this may reduce the complexity of some cases and further lower the

costs of support. In the absence of numbers to inform the analysis, we

acknowledge but do not attempt to model this effect.

While there is much uncertainty as to how alternative provision operates at

scale, the figures available to us have the advantage of being from a live

programme. We help address the uncertainty by presenting a range of

estimates.

Results

Our analysis considers three alternative forms of scaled-up provision based on

the CSP figures:

1 Estimated support cost on a day-for-day replacement basis beyond 28

days (Table 3.9)

− the lowest cost estimate, which takes the average daily cost of alternative provision and calculates the cost of replacing detention for those currently held for longer than 28 days

2 Estimated cost of year-long support beyond 28 days (Table 3.10)

− an ‘upper-central’ estimate that provides a year-long programme of provision to anyone currently held for longer than 28 days (a perhaps-strong assumption given that 95% of people spend less than six months in immigration detention and that some people may be removed from the UK after less than a year)

3 Estimated cost of year-long support to all people (Table 3.11)

− a maximum cost estimate based on everyone taking part in a year-long community alternative in lieu of immigration detention i.e. including those currently held for less than 28 days

− this is not, by itself, a reasonable cost estimate but establishes the near-absolute upper bound on the costs of a scaled-up alternative to detention

Given the above, it is reasonable to think that the additional costs incurred

would lie somewhere between the estimates from Approaches [1] and [2].

From this analysis, the lower bound estimate of alternative provision is that it

would cost around 10% of what could be saved from detaining people for no

more than 28 days i.e. around £6m each year, in the range £4-10m.

Table 3.9: Estimated cost of day replacement support beyond 28 days

Lower (£m) Central (£m) Upper (£m)

2014 3.4 - 4.4 5.3 - 6.8 7.2 - 9.2

2015 4.4 - 5.6 6.7 - 8.6 9.0 - 11.5

2016 3.3 - 4.2 5.1 - 6.5 6.9 - 8.8

2017 3.8 - 4.9 5.8 - 7.4 7.8 - 10.0

2018 3.1 - 3.9 4.7 - 6.0 6.3 - 8.0

Notes: Ranges are for the current (original) and additional (planned expansion) costs, respectively.

Source(s): Detention Action, Home Office (2019a); CE calculations.

Our analysis suggests that

the savings from reduced

detention would exceed the costs

of alternative provision



At the other end (see Table 3.10), a high level of community support would be

equivalent to around half of the cost savings: £28-37m from a central cost

saving estimate of £55-65m. At its most pessimistic, using the lower bound

cost saving figure of £35m from the previous section, the net financial impact

of the immigration detention reform could be broadly neutral. A net cost is

difficult to envisage.

A mid-point estimate (a simple average) would suggest around £15-20m or

around one-third of the cost saving.

Table 3.10: Estimated cost of year-long support beyond 28 days

Current (£m) Additional (£m)

2014 30.1 38.3

2015 35.0 44.7

2016 28.7 36.6

2017 28.5 36.4

2018 21.6 27.6


Approach [3] sets a limit on how much a scaled-up scheme might cost (see

Table 3.11). From that analysis, the results suggest that the cost of such a

programme (which we think is implausible, in any case) would come close, but

perhaps not exceed, what the Home Office has historically spent on

immigration detention. From that, it is very hard to conceive of a level of

alternative provision that could cost more than what the UK currently

spends on immigration detention. This is especially so given that the Home

Office expenditure does not represent the entirety of the costs of UK

immigration detention. It excludes, for example, escorting and healthcare

costs.

Table 3.11: Estimated cost of year-long support to all people

Current (£m) Additional (£m)

2014 82.0 104.6

2015 91.8 117.1

2016 79.2 101.1

2017 78.0 99.5

2018 70.4 89.8


Limitations

As can be seen from the results above, our analysis generates a wide range

of cost estimates for alternative provision at scale. These estimates depend

heavily on the assumptions about how the needs of the total population of

people detained compare to those of the CSP participants. That is, the

analysis is sensitive to assumptions about how a programme of alternative

provision might ultimately scale. There is little evidence on which to examine

this in depth.

Compared to the needs of the whole population, the CSP arguably provides

more support than many would need. From the earlier analysis, 65% of people

Alternative

provision could

cost up to half

the value of our

central cost

savings estimate

It is hard to see

how alternative

provision could

entirely offset the

savings from

shorter detention

lengths

The uncertainty about alternative

provision at scale leads to a

wide range of estimates



spend less than 28 days in immigration detention; 95% less than six months

and 99% less than a year. A full year of support, as is typical in the CSP, is

perhaps more than would be needed to support some people to live in the

community. Moreover, by catering for young ex-offenders, the CSP likely

provides for cases that are relatively more complex than many others. In these

respects, were the typical level of CSP provision offered to everyone, it is

likely that the estimate of the total cost would be towards the upper end.

Conversely, by not providing accommodation (as other programmes have

proposed to), it may be that the costs of alternative provision at scale may be

under-estimated in another sense.

In our lowest estimates, we assume that the equivalent of one day of CSP

support (on average) can replace one day in immigration detention, on a like-

for-like basis. This assumes that the benefits of the CSP are divisible in this

way, which is a strong assumption. At the other extreme, we assume that a

year of CSP provision would be provided to anyone who would otherwise have

spent more than 28 days in immigration detention (and would thus be released

under a 28-day time limit).8 This is a strong assumption in the other direction.

We would thus expect the cost of this alternative provision at scale to lie

somewhere between the two extremes, but it is hard to say quite where.

Eligibility and claims for financial support are difficult to estimate

Those not in immigration detention may be eligible for government support.

For asylum seekers (who represent around half of people held in immigration

detention), the principal forms of support are Section 95 and Section 4

support. Under various criteria, such as having an ongoing asylum or human

rights claim, means (‘destitution’) tests, and evidence of reasonable steps to

bring a claim, or concerns about their safety, these people may receive

accommodation and weekly payments. Support is also available in principle

for those outside of the asylum process.

Under a 28-day time limit, there would be more people outside of immigration

detention and, in turn, more people who might be eligible and go on to claim

financial support. They may also make use of certain public services (again,

under certain eligibility conditions).

It is not straightforward to cost either the financial support or the use of public

services because there is limited information on what happens to people on

release i.e. beyond the ‘outcome’ detailed earlier in this chapter (such as bail

versus removal or departure from the UK). We cannot easily know what

different people are eligible for and likely to claim; or for how long.9 Moreover,

how eligibility might change under a 28-day time limit (because it might be less

disruptive to people’s living situations etc) cannot be easily determined. This

represents a gap in our estimates of additional cost.

8 Having already established (from Table 3.11) that the most extreme version of alternative provision,

complete replacement of immigration detention with one-year support, would be no higher than existing

Home Office expenditure on detention, we do not dwell on it further.

9 The previous Matrix Evidence (2012) study developed assumptions from qualitative evidence. Discussions

with Liberty suggested that the assumption used then (about the availability of Section 4 support) is unlikely

to be tenable now owing to policy changes.



As a rough calculation, asylum support offers £37.75 per person each week

with some extra money for young children (up to £5 for a baby under one year

old).10 Following the same line of reasoning as for alternative provision, the

daily financial cost to the government could be just over £6 per person per day

(£5.39 without any child supplement). For comparison, this is at the lower end

of the per-day cost of alternative provision. On that basis, the provision of

financial support to replace days that would otherwise be spent in detention

(above the 28-day limit) would imply an additional cost of less than £5m. This

is comparable to the lower values in Table 3.9. This order of magnitude would

not be affected by whether this was provided to just asylum seekers (half the

detention population) or all people released.

However, if people were able to claim for more or longer support, the cost

would increase accordingly. We cannot be certain how much larger that cost

might be and thus how material it might prove in the context of the other costs

and savings.

We have no straightforward basis on which to calculate the impact on other

public service provision.

10 https://www.gov.uk/asylum-support

https://www.gov.uk/asylum-support



4 Conclusion

The previous chapter sets out the results from the individual elements of the

analysis. In summary:

• Holding people in detention for no more than 28 days and avoiding

compensation payments for wrongful detention could save £55-65m each

year.

− Sensitivity analysis gives a wider range of £40-90m, which sustains the argument that potential cost savings are in the tens of millions of pounds.

− Supplementary analysis produces results of similar size and, set against published expenditure figures, we can be confident in the order of magnitude of these results.

• The structure of a system of alternative provision at scale and in place of

long-term detention is much less certain. Alternatives to detention remain

at a pilot stage and their applicability to the wider population at risk is

unclear.

− Nevertheless, on a per-day basis, information from Detention Action’s Community Support Project suggests that alternative provision is substantially cheaper than immigration detention. With various caveats, it is hard to see how the additional costs incurred could outweigh the potential savings above.

− Taking the most conservative of the results we consider to be realistic, additional costs incurred could be up to £30m.

− Given the uncertainty, we have somewhat less confidence in these results and err towards the upper end of the cost range.

From the above, our central estimate is of a potential net saving of £25-

35m each year, and possibly more (see Figure 4.1). It is difficult to conceive

of a situation in which immigration detention reform could lead to substantial

net costs relative to the current system.

Figure 4.1: Savings and costs of UK immigration detention reform


Potential cost savings: £55-

65m each year

Potential new costs: up to

£30m each year

Potential net saving: at least

£25-35m each year



As we state in the previous chapter’s discussion, the main limitations of our

analysis concern:

• data quality and availability, with implications for the level of detail and

upper/lower bounds we can place on the results

− this includes how alternative provision might work at scale

• policy uncertainty, which could affect future business-as-usual trends, with

implications for the potential for future cost savings

A lack of detailed data on the operation of the immigration detention estate

(detailed costs; breakdown of all components, not just the detention centres

etc) limits what can be modelled in an analysis of this type. Insofar as we

make as much use of government data sources as we can, we have

endeavoured to produce as firm a cost saving estimate as we can.

This cost estimate is long-term only and is also narrow in scope by only

considering the operation of the detention estate. We have not been able to

estimate savings elsewhere in the system (e.g. healthcare and escorting)

leading to a possible under-estimate of the savings. Wider non-financial

benefits from, say, reduced harm to physical and mental health are also not

included. This was beyond the scope of the work, but it is important to

emphasise the fundamental non-economic debate about immigration

detention.

As we make clear throughout, there is little to go on when modelling the cost

of alternative provision at scale. Nevertheless, our approach makes use of the

available data from a live pilot programme. That pilot, Detention Action’s

Community Support Programme, is the closest we have to an indication of

how alternative provision might work. Moreover, the project has proven

promising enough for Shaw (2018) to include its expansion as one of his

recommendations to government. The programme is, however, quite specific

to the needs of a certain group of people and may provide more/longer

casework support than others might need (possibly overstating the cost).

Conversely, by not providing certain forms of support, like accommodation,

there is also the potential for an under-estimate in this respect. Overall, it

remains difficult to see how alternative provision would end up costing more

than the immigration detention system in its current form. This will become

clearer as the pilots continue.

Limitations



5 References

Detention Forum (2018) ‘Alternatives to detention: Frequently asked

questions’

http://www.detentionforum.org.uk/wp-

content/uploads/2018/08/AlternativestoDetention_A4_5.pdf

Hansard (2019a) ‘Immigration: Detainees’, Response by Caroline Nokes to

written question from Tulip Siddiq, 01/02/2019, 212416

https://www.parliament.uk/business/publications/written-questions-answers-

statements/written-question/Commons/2019-01-24/212416/

Hansard (2019b) ‘Immigration: Detainees’, Response by Caroline Nokes to



statements/written-question/Commons/2019-02-08/218833

Hansard (2019c) ‘Immigration: Detainees’, Response by Caroline Nokes to



statements/written-question/Commons/2019-02-19/223185/

Home Office (2018a) ‘User guide to Home Office immigration statistics’

https://www.gov.uk/government/publications/user-guide-to-home-office-

immigration-statistics--9

Home Office (2018b) ‘Campsfield House Immigration Removal centre to close’

https://www.gov.uk/government/news/campsfield-house-immigration-

removal-centre-to-close

Home Office (2018c) ‘Annual Report and Accounts’

https://assets.publishing.service.gov.uk/government/uploads/system/upload

s/attachment_data/file/727179/6_4360_HO_Annual_report_WEB.PDF

Home Office (2019a) ‘Immigration Statistics, year ending December 2018 –

Detention’ https://www.gov.uk/government/statistics/immigration-statistics-

year-ending-december-2018

Home Office (2019b) ‘Immigration Enforcement Transparency Data’

https://www.gov.uk/government/publications/immigration-enforcement-data-

february-2019

HM Government (2007) ‘UK Borders Act 2007‘

https://www.legislation.gov.uk/ukpga/2007/30/contents

HM Government (2019) ‘Immigration and Social Security Co-ordination (EU

Withdrawal) Bill 2017-19’

https://services.parliament.uk/Bills/2017-

19/immigrationandsocialsecuritycoordinationeuwithdrawal.html

ILPA (2011) ‘ILPA Information Sheet – Detention 2’

https://www.ilpa.org.uk/data/resources/13347/11.06-Detention-2.pdf

McGuinness, T., Gower, M. (2018) ‘Immigration detention in the UK: an

overview’ http://researchbriefings.files.parliament.uk/documents/CBP-

7294/CBP-7294.pdf









https://www.gov.uk/government/publications/user-guide-to-home-office-immigration-statistics--9

https://www.gov.uk/government/publications/user-guide-to-home-office-immigration-statistics--9

https://www.gov.uk/government/news/campsfield-house-immigration-removal-centre-to-close

https://www.gov.uk/government/news/campsfield-house-immigration-removal-centre-to-close

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/727179/6_4360_HO_Annual_report_WEB.PDF

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/727179/6_4360_HO_Annual_report_WEB.PDF



https://www.gov.uk/government/publications/immigration-enforcement-data-february-2019

https://www.gov.uk/government/publications/immigration-enforcement-data-february-2019

https://www.legislation.gov.uk/ukpga/2007/30/contents

https://services.parliament.uk/Bills/2017-19/immigrationandsocialsecuritycoordinationeuwithdrawal.html

https://services.parliament.uk/Bills/2017-19/immigrationandsocialsecuritycoordinationeuwithdrawal.html

https://www.ilpa.org.uk/data/resources/13347/11.06-Detention-2.pdf

http://researchbriefings.files.parliament.uk/documents/CBP-7294/CBP-7294.pdf

http://researchbriefings.files.parliament.uk/documents/CBP-7294/CBP-7294.pdf



Nokes, C. (2018) ‘Immigration Minister writes about alternatives to detention’,

Home Office in the media, 05/12/2018

https://homeofficemedia.blog.gov.uk/2018/12/05/immigration-minister-

writes-about-alternatives-to-detention/

Shaw, S. (2018) ‘Assessment of government progress in implementing the

report on the welfare in detention of vulnerable persons’

https://www.gov.uk/government/publications/welfare-in-detention-of-

vulnerable-persons-review-progress-report

Silverman, S. J., Griffiths, M. (2018) ‘Immigration detention in the UK’, The

Migration Observatory at the University of Oxford

https://migrationobservatory.ox.ac.uk/resources/briefings/immigration-

detention-in-the-uk/

https://homeofficemedia.blog.gov.uk/2018/12/05/immigration-minister-writes-about-alternatives-to-detention/

https://homeofficemedia.blog.gov.uk/2018/12/05/immigration-minister-writes-about-alternatives-to-detention/



https://migrationobservatory.ox.ac.uk/resources/briefings/immigration-detention-in-the-uk/

https://migrationobservatory.ox.ac.uk/resources/briefings/immigration-detention-in-the-uk/


Cambridge Econometrics

Acknowledgements

This report was commissioned by Liberty.

The report was supported by Barrow Cadbury

Trust and the A B Charitable Trust.

Barrow Cadbury Trust is an independent charitable

foundation, committed to bringing about socially just

change.

The A B Charitable Trust (ABCT) is an independent,

UK-based grant-making organisation that is

concerned with promoting and defending human

dignity.

Documents

Economic impacts of immigration detention reform - Cambridge … · 2019. 5. 8. · Economic impacts of immigration detention reform Cambridge Econometrics 3 Authorisation and version